ML20063P079

Executive Summary Rept on Adequacy of Permanent Hydrogen Mitigation Sys for Sequoyah Nuclear Plant

ML20063P079
Person / Time
Site:	Sequoyah
Issue date:	09/30/1982
From:	TENNESSEE VALLEY AUTHORITY
To:
Shared Package
ML20063P076	List: ML20063P075 ML20063P079
References
NUDOCS 8210120289
Download: ML20063P079 (31)

Text

ENCLOSURE EXECUTIVE

SUMMARY

REPORT ON IBE ADEQUACY OF THE PERMANENT HYDROGEN MITIGATION SYSTEM POR THE SEQUOYAH NUCLEAR PLANT SEPTE8EBER 1982 TENNESSEE VA11EY ADIBORITY t

82101202M 27

PDR ADOCK O PDR P _

. I. ' Introduction This report is an executive summary whose purpose is to provide an overview of the Tennessee Valley Anthority's (TVA) position that the Permanent Hydrogen Mitigation Systan (PHMS) is an adequate hydtcson control system for the Segnoyah Nuclear Plant and would perform its intended function in a manner that provides adequate safety margins.

Highlights of the PENS design and supporting analyses and research are presented. A more comprehensive technical summary is provided as an attachment to this report.

I l

- ' ~ ' - - - - - - --- , . _ , , , . . _ , _

, ~ . . .

II. Per===ent Hydromen Mitimation System (PEMS) Descrintion TVA has selected the concept of controlled ignition using thermal i ignitors for the PHMS at the Sequoyah Nuclear Plant. Briefly, the concept is to reliably ignite loan hydrogen-air mixtures throughout the containment to achieve periodic or continuous burning. This

, moderated energy addition rate would allow the contaissent heat sinks to absorb the heat of combustion more effectively and reduce the overall containment pressurization. This selection was made af ter a number of alternatives were thoroughly evaluated.

In early 1980, the TVA Board of Directors requested the TVA staff to investigate potential. mitigation systems for degraded core accidents at Sequoyah. An intensive study was undertaken of concepts to prevent or minimize the effects of hydrogen combustion as well as concepts to increase containment capacity for overpressure events.

Af ter evaluating each of these strategies, the TVA staff recommended the implementation of a controlled ignition system. This concept was the basis for the Interim Distributed Ignition System (IDIS) installed at Sequoyah in the summer of 1980. Beyond this commitment to the IDIS, TVA, tcgether with Duke Power and American Electric Power ( AEP), continued to investigate alternative methods of hydrogen contr ol . Af ter completing these evaluations and camparing the alternatives, TVA selected controlled ignition for the PENS.

A durable thermal igniter capable of maintaining an adequate surf ace temperature was specified for the PRMS. An igniter developed by Tayco Engineering to operate at a standard plant voltage of 120V ac was selected and has been shown to be capable of maintaining an adequate surf ace temperature for extended periods, initiating i combustion, and continuing to operate in various combustion environments. To assure adequate coverage, a total of 64 ignitors will be distributed throughout the major regions of containment in which hydrogen could be released or to which it could flow in significant quantities (see figure in attachment). There will be at least two igaiters, controlled and powered redundantly, located in each of these regions.

i The PENS components inside containment will maintain their functional capability under the effects of postaccident conditions including combustion. In addition, the PBMS camponents will be seismically supported.

The ignitors in the PENS are equally divided into two redundant '

groups to ensure adequate coverage even in the event of a single failure. Manual control and status indication of each group will be provided in the main control room. The system would be energized manually f ollowing the start of any accident which indicates inadequate core cooling without waiting for any hydrogen buildup.

! Separate trains of power will be provided for each group of ignitors and will be backed by automatic loading onto the diesel generators upon loss of of f site power.

In addition, appropriate surveillance testing requirements and technical specifications have been provided.

?

b We conclude that the pelts design, as described here, is adequate and that the system wonid perform its intended function in a manner that provides adequate safety margins.

l

'III. Sunnortina Analyses Numerous analyses have been performed by TVA and its subcontractors

-during the past two years to study the effects of mitigating hydrogen by controlled ignition on ice condenser containment structures and equipment during selected degraded core accidents.

Calculations of containment atmospheric pressure and temperature have j been performed using the CLASIX computer code developed by Westinghouse Off shore Power Systems. The CLASIX code results have be'en compared f avorably to results from other containment codes. The code also has been shvwn to conservatively predict the response f rom several experiments. For input to the CLASIX code, values for combustion parameters were obtained f rom the literature and values for hydrogen and steam release rates were calculated with the NRC-funded MARCH code. Enough sensitivity studies were performed on containment parameters, combu* tion parameters, and release rates to reasonably bound the expecteo esponse. The calculated peak containment pressure for the base case set of parameters was 19 psig while the highest pressure calculated in the sensitivity studies was less than 28 psig.

The response of the containment shell and internal structures to these static pressure loads has been evaluated. The minimum calculated structural capacity at yield of 45 psig bounds these calculated internal pressures with considerable margin.

Our analyses and research have indicated that dynamic loads f rom a detonation do not have to be considered because detonation is not a credible phenomenon in the containment. Briefly, this is because:

i (a) there are no high-energy sources to initiate a detonation, i (b) there would be no rich concentrations throughout the containment because the distributed ignitors would initiate combustion as the mixture reached the lower flammability limit and because effective mixing would occur, and (c) there are no areas of the containment i

I with sufficient geometrical confinement to allow for the flame acceleration necessary to yield a transition to detonation. H ow ev er, at the NRC's request, TVA has calculated the responso of the l containment shell to an impulse pressure from a hypothetical local detonation. The results showed that a margin of saf ety of three existed before material yield would be reached.

The survivability of key equipment has been evaluated for the calculated atmospheric pressure and temperature profiles augmented by radiative fisme effects. The equipment temperature response was calculated using the NRC-funded HEATING 5 code and the results were compared with the original qualification temperatures. This comparison showed that the key equipment would survive under postaccident conditions including combustion. .

In summary, these analyses have demonstrated that the containment structures and key equipment would survive the effects of selected degraded core accidents when mitigated by the PHMS and continue to remain intact and operational. We conclude that the PENS, as supported by the analyses described here, is adequate and would perform its intended function in a manner that provides adequate saf ety margins.

,7

, JV.* Sunnortina Research Extensive research has been sponsored by TVA, Duke, AEP, and Electric Power Research Institute (EPRI) during the past two years to study hydrogen combustion, distribution, and mitigation. The research programs were designed to be confirmatory in nature. - They were necessarily limited in scope and depth due to time constraints imposed by the Sequoyah operating license conditions and the availability of test f acilities. The progr ams f ocused on the engineering applications of hydrogen combustion technology in support of a mitigation system.

TVA, Duke, and AEP spor.sored combustion experiments at Fenwal Incorporated to investigate the ignition characteristics and reliability of the General Motors (GM) igniter used in the Interin Distributed Ignition System. TVA, Duke, AEP, and the EPRI sponsored an integrated research program at Whitoshe11 Nuclear Research Establishment, Factory Mutual Research Corporation, Acurex Corporation, and Hanford Engineering Development Laboratory. In one phase of the Whitoshe11 tests, the loan inition limits and minimum surf ace temperatures were determined for Loth the GM and Tayco ignitor. In other tests at Whitoshell, the extent of reaction of lean mixtures, the behavior of deflagrations in rich mixtures, the effects of fan- and obstacle-induced turbulence, and the behavior in an extended vessel geometry were each investigated. At Factory Mutual, the pressure suppression effects of a water micro-fog were studied in small scale. In the intermediate-scale testa at Acurex, the effects of' igniter location within the test vessel and the presence of a watei micro-fog were both investigated. Simulation of postaccident conditions in en ice condenser lower compartment was performed at Hanford to study the potential for hydrogen pocketing or nonuniform distribution. TVA also conducted experiments at its Singleton Laboratory on the survivability of electrical cables and the durability of ignitors under cycling, endurance, and combustion conditions.

The original research programs have been successfully concluded and the data have been submitted to the NRC. The tests showed no unexpected resul ts and confirmed the judgments made in the design and analysis supporting the PENS. Both types of ignitors were shown to be reliable and ef fective under a wide range of conditions. In general, the combustion parameter results agreed with values from the literature. In particular, the transient tests exhibited asquential combustion accompanied by relatively mild pressure rises which are characteristic of the behavior calculated with the Q,ASIX code. No detonations were ever observed even at high concentrations of hydrogen or in an extended vessel geometry. The micro-fog was ineffective as a heat sink for pressure suppression during t combustion. The Hanford simulation showed good mixing with no pocketing of hydrogen.

We conclude that the PRMS, as supported by the research here, is adequate and would perform its intended function in a manner that provides adequate safety margins.

V. Conclusions ,

TVA has designed a Permanent Hydrogen Mitigation System employing controlled ignition to mitigate the effects of hydrogen during .

potential degraded core accidents at the Sequoyah Nuclear Plant. The system is redundant, capable of functioning in a postaccident environment, seismically supported, capable of actuation from the main control room, and has an emple number of ignitors distributed throughout the contaiment. The containment structures and key equipment have been shown by analysis or testing to survive the pressure and temperature loads f rom selected degraded core accidents and to continue to function. An extensive research program has confirmed our analytical assumptions, demonstrated equipment survivability and shown that controlled ignition can indeed mitigate the effects of hydrogen releases in closed vessels. We conclude that the PENS is an adequate hydrogen control system that would perform its intended function in a manner that provides adequr te safety margins.

f

. . _ . . _ , _ . - . - - _ . ,.m __ ..,.-- -, - - - - -

l I

e ATTAGNENT 10 ENCLO6URE TECHNICAL

SUMMARY

REP 0itT ON THE ADBQUACY OF THE PERMANENT HYDROGEN MITIGATION SYSTEM FOR H E SBQUOYAH NUCLEAR PLANT SEPTEISER 1982 TENNESSEE VALLEY AU1EORITY l

i

. TABLE (EF CONTENTS ,

. i I. Introduction II. Permanent Hydrogen Mitigation System Description III. Supporting Analyses A. Structures B. Equipment IV. Supporting Research A. Fenwal Incorporated B. Whiteshell Nuclear Research Establishment C. Factory Mutual Research Corporation .

D. Acurez Corporation E. Hanford Engineering Development Laboratory F. TVA Singleton Materials Engineering Laboratory V. Conclusions VI. References

+

l l

l l

l l

.~,s..

I

I. Introduction This report is a technical summary whose purpose is to substantiate the Tennessee Valley Authority's (TVA) position that the Permanent Hydrogen Mitigation System (PHMS) is an adequate hydrogen control system for the Sequoyah Nuclear Plant and would perform its intended function in a manner that provides adequate saf ety margins. The report draws from and ref erences the many technical reports that have been submitted by TVA to the NRC over the past two years. First, the critoria and final design for the PENS is described. Next, a discussion is provided of the numerous analyses performed to determine the effects on key structures and equipment of mitigating degraded core accidents with the PENS.

Last, the research program conducted to confirm our understanding of hydrogen combustion control is reviewed. Throughout this ,

report, resolution of the various technical issues that have been raised (contsinnent capability, equipment survivablity, local detonation, etc.) is provided and application of the test data and analyses is made in support of the adequacy of the PENS.

_. = __. - _ _ .- --.

,II . Per===ent Hwdroman Mitimation Swaten (PRMS) Descrintion TVA has selected the concept of controlled ignition using thermal ignitors for the PENS at the Sequoyah Nuclear Plant. Briefly, the concept is to reliably ignite loan hydrogen-air mixtures throughout the containment to achieve periodic or continuous burning. This moderated energy addition rate would allow the containment heat sinks to absorb the heat of combustion more effectively and reduce the overall containment pressurization.

This selection was made af ter a number of alternative concepts were thoroughly evaluated and compared. In early 1980, . the TVA Board of Directors requested the TVA staff to investigate potential mitigation systems for degraded core accidents at S equoy ah. An intensive study was undertaken of concepts to prevent or minimize the effects of hydrogen combustion such as preinerting with nitrogen, postinerting with Halon, or controlled ignition. Also investigated were concepts to increase containment capacity for verpressure events such as augmented atmospheric cooling or various forms of containment venting. Each of these J

mitigation strategies was evaluated based on their ef fectiveness, technical f easibility, additional risk, reliability, and cost.

The report recommended the implementation of a controlled ignition -

sy s t em. This concept was the basis for the Interim Distribution Ignition Syst em (IDIS), installed at Sequoyah in the summer of 1980.

Beyond this commitment to the IDIS, TVA, together with Duke Power and American Electric Power ( AEP), continued to investigate l

alternative methods of hydrogen control. The potential electromagnetic interference effects of spark igniters were examine d. A conceptual design study for a postaccident Halon 1301 injection system was commissioned. The corrosive effects on stainless stoel of Halon decomposition products were later danonstrated by TVA at its Singleton Materials Engineering Laboratory. Beach-scale tests on controlled combustion with catalytic combustors were performed and the effects of catalyst poisoning by fission products were investigated. TVA also evaluated controlled ignition enhanced with spray fogging, oxygen removal with a gas turbine, and postaccident inerting with carbon dioxide. Af ter completing all these evaluations and camparing the.

i alternatives, TVA selected controlled ignition for the PENS.

l Brief descriptions are provided below of the PBMS and its design criteria, operating procedure, surveillance testing, and technical j specifications.

l l To assure that hydrogen would be ignited at any containment 1ccation as soon as the concentration exceeded the lower flammability limit, a durable thermal igniter capable of i' saintrining an adequate surface temperature was specified. An igniter developed by Tayco Engineering was selected for use in the l PENS since it operates at a more standard plant voltage aof 120V ac than the lower voltage required by the General Motors (GM) slow plus used in the IDIS at Sequoyah. The Tayco model igniter has been shown by experiment to be capable of maintaining surf ace temperatures in excess of the required minimum for extended periods, initiating combustion, and continuing to operate in

vericas combustica savircamssts. Infermatica cm stah proci

. lesting is included in sections IV.B and IV.F of this summary r e por t. ,

I To assure adequate spatial coverage, a total of 64 igniters will ,

be distributed throughout the major regions of containment in l which hydrogen could be released or to which it could flow in significant quantities (see figure) . There will be at least two '

ignitors, centro 11ed and powered redundantly, located in each of these regions. Following a degraded core accident, any hydrogen which is produced would be released into the lower compartment inside the crane wall. To cover this region, 22 igniters (equally divided between tr3 1 ns) will be provided. Eight of these will be distributed on the reactor cavity wall exterior and crane wall interior at an intermediate elevation to allow the partial burning that accompanies upward flame propagation. Two ignitors will be located at the lorer edge of each of the five steam generator sad pressurizer enclosures, two in the top of the pressurizer enclosure, and another pair above the reactor vessel in the cavity. These 22 lower compartment ignitors would prevent flammable mixtures f rom entering the ice condenser. Any hydrogen not burned in the lower campartment would be carried up through i the ice condenser and into its upper plenum. Since steam would be removed from the mixture as it passed through the ice bed, thus concentrating the hydrogen, mixtures that were nonflammable in the lower compartment would tend to become flammable in the ice condenser upper plenum. This phenomenon is supported by the CLASIX containment analysis code (discussed in section III. A of this sammary report) which predicts more sequential burns to occur in the upper plenum than in any other region. Controlled burning in the upper plenum is preferable since the amount of hydrogen consumed in each lean-limit burn is so low due to the relatively sus 11 volume of the region that the energy addition rate to the containment is moderated. We also conclude, based on the expert opinion of Dr. Bernard Lewis and Bela Karlovitz, that there is no realistic potential for a transition to detonation in the upper plenum because the available i ignition strength is weak, the entering mixtures will be just-l flammable, and the plenum does not have sufficient geometrical confinement above or below the region of cambustion. Therefore, we have chosen to take advantage of the beneficial combustion characteristics of the upper plenum by distributing 16 ignitors equally around it. Four ignitors will be located around the upper compartment dome, four more around the top inside of the crane wall, and one above each of the two air return f ans. The air return f ans provide recirculation flow from the upper compartment through the ' dead-ended' volume and back into the main part of the lower compar tment. To cover this region, there will be a pair of ignitors in each of the rooms (a total of 16 ignitors) through which the recirculation flow passes.

The PENS components inside containment will maints.in their functional capability under postaccident conditions. These components will survive the effects of multiple hydrogen burns and will be protected from spray impingement and flooding. In addition, the PENS components will be seismically supported.

l

The ignitors in the PEMS are equally dividsd into two redundant groups, each with independent and separate controls, power, and locations, to ensure adequate cover >ge even in the event of a single f ailure. Manual control of each group of igniters will be provided in the main control room and the status (on-off) of each group will be indicated there. The system would be energized manually following any accident upon the occurrence of any condition which indicates inadequate core cooling without waiting for a potential hydrogen buildup. Separate trains of Class 1E 480V ac auxiliary power will be provided for each group of ignitors and will be backed by automatic loading onto the diesel generators upon loss of off site power. Each individual circuit will power two ignitors and have a design vcitage of 120V ac. .

Surveillance testing proposed for the PHMS will consist of energizing the system from the main control room and taking voltage and current readings from each circuit at the distribution panela located in the anziliary building. These readings can then be compared to ones taken during preoperational testing of the system to indicate whether or not both ignitors on each circrit

~

are operational without requiring containment entry. The operability of at least 31 of the 32 ignitors per train would conservatively guarantee an effective coverage throughout the containment. Appropriate technical specifications on test intervals and restoration to operable status have previously been pr opo se d.

We conclude that the PHMS design, as described here, with igniter type and locations, redundancy, capability of functioning in a postaccident environment, seismic support, main control room actuation, and remote surveillsace is adequate and the system would perf orm its intended function in a manner that provides adequate saf ety margins.

III. Sannortina Analyses Numerous analyses have been performed by TVA and its contractors during the past two years to study the effects of mitigating '

hydrogen by controlled ignition on ice condenser containment structures and equipment during selected degraded core accidents.

Calculations of containment atmospheric pressure and temperature during these accidents have been performed using the CLASIX code.

The response of the containment shell and internal structures to the peak calculated pressures has been evaluated. The response of the contaimaant shell to an impulse pressure f rom a hypothetical local detonation has been calculated. The survivability of key equipment has been evaluated for the calculated atmospheric pressure and temperature profiles augmented by radiative flame effects. The analyses have demonstrated that the containment structures and key equipment would survivo the effects of selected degraded core accidents when mitigated by the PRMS and continue to remain intact and operational. We conclude that the PENS, as supported by the analyses described below, is adequate and would perform its intended function in a manner that provides adequate saf ety margins.

A. Structures Containment atmospheric pressure loadings on the shell and internal structures during degraded core accidents including hydrogen combustion have been calculated using the OLASIX containment analysis code written by Off shore Power Systems (OPS),

a division of Westinghouse. The expertise developed over the years in writing and verifying NRC:-accepted design basis containment analysis codes was used as a basis f or this ef fort.

The ice condenser containment was modeled in OLASIX using such standard assumptions as homogeneous volume rodes. Extensions to this traditional methodology were included in the code to account for the eff ects of degraded core accidents such as hydrogen combustion. Hydrogen combustion was represented by a simple model that added the heat released during burning to the surroundings when flammability criteria were met in that region. The CLASIX code has been compared by OPS to IMD, an NRC-accept ed subcampartment ice condenser analysis code, and to 0000 CLASS 9, a degraded core accident containment analysis code based on the NRC-accepted 0000 code. The comparisons showed good agreement. The CLASII code was also used to model hydrogen combustion experiments conducted at Fenwal Incorporated and Lawrence Livermore National Laboratory. The code conservatively overpredicted the pressure and temperature response measured during the tests. We conclude that the CLASIX code is adequate to use f or conservative prediction of the ice condenser containment response to degraded core accidents including hydrogen combustion.

The CLASIX input required to model the Sequoyah containment response to such an event consisted largely of physical parameters such as volumes, areas, and material properties that have been used previously in design basis licensing analyses. Several of these permasters, including containment spray flow rate, initial ice mass, and air return f an flow rate, s cre varied in sensitivity

- - -,.,,_.---m-- - - _ _ -

. , . _ , , y , . - - - _ _ - , , - . - . . _ , .- --.

st udi e s. In addition, several hydrogen comb 3stica parasotars vero specifiable in the input to allow for a wide range of sensitivity st udi e s. These include the lower flammability limit (LFL), the f raction of burn campleteness, and the burn duration. The burn duration actually represents the pressure rise time based on flame propagation at a constant speed af ter simultaneous ignition at all ignitors located in that volume. In our latest studies, the conservative assumptions used in the base case calculation were an LFL of 8 volume percent, a burn fraction of 85 percent, and a fleme propagation speed of 6 f t/sec. The parameters assumed in the best estimate calculation were an LFL of 6 volume percent, a burs fraction of 60 percent, and a propagation speed of 3 f t/sec.

• In the various sensitivity studies, the LFL was varied betweea 4 and 10 volume percent, the burn completeness fraction between 40 and 100 percent, and the burn duration based on flame speeds between 1 and 12 f t/sec. These value ranges are supported by numerous ref erences in the literature for turbulent combustion in loan-limit mixture s. Results from the recent Electric Power

Research Institute (EPRI) -atility lean-limit hydrogen combustion experhaents validated the use of these value ranges. Inf orma tion and conclusions f rom this combustion research is included in sections IV.B. IV.C. and IV.D of this summary report. In further cmaparisons to actual data, as stated above, the CLASIX code was able to conservatively overpredict experimental pressures measured at two different f acilities. The parameter sensitivity studies were performed to bound reported data and to account for such postulated phenomena as steam inerting the lower compartment or fogging reducing the burn completeness in the upper plenum. We conclude that the cambustion parameter input, including sensitivity variations, is adequate to be used in the OLASIX code for conservative prediction of containment response.

Another set of CLASIX input parameters required to model a degraded core event included the hydrogen and steam release rates into the containment. Allowances were made in the CLASIX code for these input parameters to be varied over a wide range since they would be dependent on the accident sequence being studied. A maall-break LOCA with f ailure of saf ety inj ection (S D) 2 was chosen as the base case for analysis because it is similar to the TMI-2 class of accidents. The S,D event is also an appropriate selection because it is believed to Be the most probable accident sequesce that would result in core damage at Sequoyah. Recovery of core cooling was assumed to occur prior to core slump and the cladding reaction was terminated at a conservative level of 75 percent. In addition, a review of other probable scenarios shows the S D 2 transient results in more than twice as much hydrogen generation prior to core slump as was found in the other scenarios. Beyond the S2 D base case, sensitivity studies were perfonaed to evaluate the eff ects of increasing the hydrogen release rate throughout the event by as much as a factor of three and increasing the rate in a

' spike' feshion over a segnert of the event. In addition, the hydrogen release rates from analyses (using the MARCH camputer code) of a number of other accident sequences were reviewed and found to be bounded by either the S 2D base case or the sensitivity studie s. The S D base case release rate used in the TVA analysis 2

also bounded the release rates presented in NURBG/CR-2540, ' A 1

Method fer the Analysis of Hydreass and Stcan Estesses to

, Containment During Degraded Core Cooling Accidents.' Since the PHMS is intended to mitigate degraded oore events which are terminated prior to core slump, the rolesse rates de:ing the core recceery phase were calculated and also found to'be less than already covered by the studies. We conclude that the hydrogen and steam release rate input, including sensitivity variations, is adequate to use in the OLASIX eode for conservative prediction of containment response.

~

The CLASIX code calculations for the base case set of input parmasters described above resnited in a peak containment pressure of 19 psig. The best estimate case resnited in a peak pressure of less than 12 pais, the containment design pressure. The highest peak pressure that resnited from any of the numerous sensitivity studies was less than 28 psig. As described below, the Sequoyah containment yield strength has been calculated to be at least 45 pois.

i Structural analyses have been perf ormed to determine the static pressure capability of the containment and internal struotares.

The pressure rise resniting f rom a hydrogen deflagration is slow enough to be treated as a static pressure load in the analysis.

The associated temperature ef fects were f ound to be negligible.

An elastic-plastic analysis was perfossed by TVA using a finite element model of the limiting section (1/2' cylindrical plate between elevations 756' 3' and 810' 3') of the steel containment shell.- All other containment boundary components were evaluated and it was determined that this shell section was limiting in terms of containment yield strength. Using the actual minimum

yield strength of the plate material, the yield pressure of this shell section was found to be at least 45 psig. Other independent structural evaluations have been made that confirmed this minimum

! capa ci ty. An evaluation was also made of the concreto divider deck (the main internal struetare between the upper and lower compartment) that revealed its differential pressure capacity to be equal to or greater than the containment shell capacity. We ,

conclude that the capability of the contaimaant shell and internal structures is adequate to withstand the static pressure loads during hydrogen combustion in the degraded core accidents I studied.

l l In addition to these analyses of static pressure capability, TVA has performed an analysis of the dynamic response of the containment to an impulse load from a hypothetical local l detonation. Development of the bapalse load and the structural analysis was requested by the NBC, although our analyses and research have indicated that local detonation is not a credible l phonamenon in the containment. To briefly review, several factors affect the po'tential for a detonation including ignition strength, hydrogen concentration, and geometrical confinement. Addressing l these factors individually, the thermal ignitors used f or

! controlled ignition are.considerod by experts, including Dr. Roger Strehlow (an NRC consul tant), to be 'sof t' or ' weak' sources of ignition and as such are .not likely initiators of detonation.

Second, rich coneantrations of hydrogen will not be present ,

.. ~ . - - - - - ---- ,,.,-g -r- .e-.,. --- - - . - , - - - - - - - - - . - - - > ~ - , , , . ---,-,nw,

i 4

throughout large regions of the containment because the PEMS ignitors will initiate combustion near the LFL. h is has been demonstrated on numerous occasions (see sections IV. A, IV.B, IV.C, and IV.D) including' tests in the presence of steam or spray. In addition, isolated rich concentrations away from the source due to extreme hydrogen gradients or pocketing will not occur. This has been confirmed by results from the mixing tests in the simulated ice condenser containment at Manf ord Engineering Developent Laboratory (see section IV.E) . Third, we have identified no areas of the containment with sufficient geometrical confinement to allow for the extreme flame ace >1eration necessary to yield a transition to detonation. For example, the vertical ice baskets in the ice condenser are not sufficiently confined radially and the circumf erentist apper plenum above the ice condenser is not sufficiently confined above or beltw for a transition to detonation to occur (see section II). Even if rich mixtures were postulated to exist in a confined geometry, it is improbable that a detonation would result. Illustrating this f act are two of the tests conducted at Whiteshell Nuclear Research Establishment that f ailed to produce a detonation when igniting a stoichiometric (about 29.5 volume percent hydrogen) mixture in an enclosed sphere or even when igniting a 25 volume percent mixture in a pipe attached to the sphere in a configuration more conducive to a transition to detonation. For more information see section IV.B of this summary report. We conclude that detonation is not a credible phencuenon in the ice condenser containment. However, as stated above, TVA has developed an impulse load from a hypothetical local detonation and analyzed the dynamic containment i response. The hypothetical load was based on the detonation of a six-foot diameter spherical cloud with wave speeds (to calculate the pressure rise time) and peak overpressures obtained from the literature. The impulse was assumed to act at the center of the same critical containment shell sootion used for the static i analysis. The results showed that a margin of safety of three existed before material yield would be reached. We conclude that the containment shell could survive even such a hypothetical local detonation.

Based on the above analyses, we conclude that the containment structures would survive the effects of selected degraded core accidents when mitigated by the PEMS and continue to remain intact.

B. Eauipment Containment atmospheric pressure and temperature loadings on key equipent in the containment have been calculated using the GASIX code discussed above in section III. A. The parameters assumed previously for the base case were used again except that the burn duration was based on a low flame speed of one f t/sec chosen at the NRC's request to enhance the heat contribution from the fi sm o. To account for these flame effects, the EASIX temperature transient in each of the regions containing key equipment selected for analysis was augmented by a radiative heat fluz term. The radiative host flux was imposed during each burn agd was based on a conservative adiabatic flame temperature of 1400 F. This T

l

combined temperature load was impos:d on the cqalpaczt la ca

. analysis using the standard BEATINGS thermal code which was developed with NRC f unding. The equipment was initially assumed to be in equilibrium at the highest proburn atmospheric temperature resulting f rom the postulated degraded core accident.

The thermal analysis was extended until well af ter all the temperature peaks associated with barns had passed.

Key equipment inside containment essential for safe shutdown of the plant was identified. That subset of equipment either considered to be potentially sensitive to temperature or located in regions of numerous burns such as the ice condenser upper plenum was then selected. This subset would bound the remaining key equipment items for the evaluation of temperature survivability. The pressure capability of the key equipment was judged to be controlled by the limiting containment shell section pressure capability described above in section III. A. The subset cf key equipment included the exposed incore thermocouple cable and hot and cold leg RTD cable, the Interin Distributed Ignition

' Syst em (IDIS) igniter assembly, the igniter assembly power cable in conduit, and a transmitter assembly representative of the types installed in the plants. The decision was made to test the exposed cables rather than attempting to analyze them due to the potential for changing surf ace properties (see section IV.F).

Thermal analyses were performed on the remaining key components.

The ignitor assembly analysis was perf ormed on a Sequoyah IDIS assembly which should conservatively bound the PRMS assembly response. It showed that the coge of the transformer inside the igniter assembly would reach 1577 whjle the transformer windings were designed to operate at up to 428 F. Analysis also showed thal the conduit f or the igniter assembly power cable would reach 3327 (and the interior even less) while tests conducted at TVA's Singleton Laboratory showed e cable in conduit would function without degradation up to 600 . The transmittLr analysis resulted in a casing surface temperature of 2457 (and the interior even gess) while the transmitter has been qualified to operate at 320 F. This thermal analysis methodology was compared I to an NRC-accepted Westinghouse equipment thermal qualification model and showed good agreement. In addition, the methodology was applied to sample Fenwal test data and found to conservatively overpredict thermal response.

In addition to the key subset described above, the effects of temperature and pressure were evaluated for other key equipment such as the air return f ans. No burns were predicted by CLASIX to occur in the upper compartment for the base case parameter assumptions. How ever, even for those sensitivity studies which resulted in upper compartment burns, the atmosphere only very briefly exceeded the elevated temperatures at which the f ans were designed to operate in an emergency. In addition, the massive f an motor and casing (weighing approximately 1300 lbs.) have a significant amount of thermal inertia. The backdraf t dampers above the f ans avoid pressure loads on the f ans during lower campartment pressurization. Again, no upper compartment burns are predicted for the base case. How ev er, the f an blades have been

structurally analyzed to take a static load (in additica to the normal operating stresses) greater than even the maximum peak dif f erential pressure predicted in the sensitivity studies discussed in section III. A.

In addition to analyzing the survivability of the key equipment described above, special areas such as the f oam insulation around the ice condenser were evaluated for temperature effects. A thermal analysis using the HEATING code mentioned above was performed by Duke Power to evaluate whether heat f rom combustion in the ice condenser could decompose the foam to form flassable products. The analysis showed that even the heat flux f rom a constant band of fisme applied locally for 45 minutes to the ice condenser walls would not be sufficient to elevate the f oam behind it to its pyrolysis temperature.

Based on the above analyses and tests, we conclude that the contaissent key equipment would survive the effects of selscted degraded core accidents when mitigated by the PBitS and continue to remain operational.

t 9

' Il d % 'g ;

T

- - - ~ -- . . , , . , ,

- - . .. - _ . ~ . .

IV. Sansortina Research Extensive research has been sponsored by IVA, Duke, AEP, and EPRI

during the past two years to study hydrogen combustion, mitigation, and distribution. The research programs were designed  !

to be confirmatory in nature. They were necessarily limited in l scope and depth due to time constraints imposed by the Sequoyah operating license conditions and the availability of test facilities. The programs focnded on the engineering applications of hydrogen combustion technology in support of a mitigation sy st em. TVA, Duke, and AEP sponsored combustion experiments at Fenwal Incorporated. TVA. Duke, ARP, and EPRI sponsored research at Whitoshe11 Naelear Research Establishment in combustica and igniter development, at Factory Mutual Research Corporation in combustion and mitigation, at Aearez Corporation in. combustion and mitigation, and at Hanford Engineering Development Laboratory in

, distribution. TVA conducted experiments at its Singleton Laboratory in equipment survivability and igniter development. j I The original research programs have been successfully concluded and the data have been submitted to the NRC. To summarise, the tests showed no unexpected results and confirmed the judgments made in the design and analysis supporting the PEMS. We conclude that the PBMS, as supported by the research described here, is adequate and would perform its intended function in a manner that provides adequate safety margins.

A. Ianiter Performance Testina - Fenwal. Incornorated A two phase experimental program was madertaken at Fenwal to investigate the ignition characteristics and reliability of the General Motors (GM) igniter. The test vessel was a 134 f t steel sphere that was heated and insulated. Phase 1 consisted of a series of premixed combustion tests with hydrogen concentrations at 8,10, and 12 volume percent. The effects of fan-indsoed turbulence and steem addition were investigated in several tests.

l The performance of the GM igniter in igniting hydrogen mixtures was demonstrated to be reliable. In addition, comparison of such test results as pressure rises and ignition limits with previously published information showed good agreement.

The Phase 2 follow-on tests consisted of further premixed tests with hydrogen concentrations between 5-10 volume percent, tests where hydrogen was continuously inj ected into the test vessel, and a series of tests using water sprays. The most Laportant result j of the Phase 2. program was the ability of the igniter to reliably ignite lean hydrogen mixtures under adverse conditions, including the presence of steam and water sprays, and to continue to operate. The minimal pressure rises experienced during the continuous inj ection tests indicated the igniter's capability to initiate local combustion of hydrogen-air mixtures just as they became flammable. The series of sequential barns that occurred during the continuous injection tests were characteristic of the behavior predicted with the CLASIX code (section III. A) . No I

detonations were ever observed even when pure hydrogen was being admitted to the vessel during the transient tests.

I i

1

,B. Hydrogen Combustion Phenomena - Whitoshe11 Nuclear Research Establishment The experimental program at Whitoshe11 consisted of a maall-scale igniter testing segment and a multif aceted large scale segment abned at enhancing our understanding of basic combustion phenomena. The results of this program are sammarized below.

Small-scale tests were perf ormed in a 17-liter vessel to provide l further evidence of the capability of both GM and Tayco thermal '

ignitors to reliably ignite loan hydrogen mixtures. Numerous tests were conducted to determins the lower ignition limits and corresponding igniter surface temperatures in various premixed hydrogen-air-steam mixtures. Bydrogen concentrations were varied between 4-15 volume percent and steen concentrations varied between 0-60 vcluso percent. The measurement of igniter surf ace temperature required for ignition showed that the igniter at its normal operating temperature has considerable margin even f or high steam concentrations.

The larger-scale tests were performed in the Whiteshe11 Containment Te st Facility using a 223 f t3 heated and insulated metal sphere and, for some tests, a 20-foot long by 1-foot diameter attached pipo. These tests were grouped into four

, principal areas:

(a) Extent of reaction of loan mixtures (b) Laminar spherical deflagration (c) Ef f ects of f an- and obstacle-induced turbulence (d) Extended geanetry (sphere and attached pipe)

The lean mixture tests were performed in the sphere to investigate the extent of reaction under various conditions of steam and f an-induced turbulence. Hydrogen concentrations were varied between 5-i 11 volume percent and steam between 0-30 voinme percent. Fans l

were activated in several of the tests. Results were in agreement i

with previously-published data on the flammability of lean mixtures. Results also showed that the addition of relatively i

large (over 30 volume percent) amounts of steam reduced the pressure rise following burns due to the added heat capacity.

This indicates that pressure rise data from dry tests may be overconservative f or application to plant environments with high steen concentrations. Results also showed that turbulence increased the rate and magnitude of p-essure rise f or a given concentration by increasing the born completeness, thus corroborating the Fenwal results. This indicates that burning at relatively lean concentrations would be promoted by the turbulent plant conditions.

The Imminar spherical deflagration tests were performed in the sphere to compare the actual pressure rises with the corresponding theoretical adiabatic pressure rises and to confirm that no detonations would result even at high concentrations of hydrogen.

Hydrogen concentrations were varied between 10-42 volume percent and steam between 0-40 volume percent. Fans were activated in several tests. Results again showed that the addition of large amounts of steam reduced the pressure rise following burns. The l

,it

- - , , - , , . , _ , , n, _ ,~, - . .-,  ! - - , -. - - . , - . . - - - , - - - - - - - - - - - - - - - - - - _ - - - - - -

actual presstro wcs oletys Icso then the thocratiosi prosstro cad

. the margin increased .as the hydrogen concentration was increased.

No detonations were observed even at stoichiometric and higher concentrations of hydrogen which are classically considered to be detonable.

The turbulence tests were performed in the sphere to investigate the effects of turbulence induced by f ans and gratings on the extent and rate of combustion. In these tests, hydrogen concentrations varied between 6-27 volume percent. One test was run with 10 volume percent stess. Results showed that for rich mixtures, forced turbulence did not increase the overall pressure rise but did increase the rise rate slightly. In lean mixtures without f ans, the presence of gratings tended to increase the magnitude and rate of pressure rise. At high concentrations or with f ans, the gratings reduced both the magnitude and rate of pressure rise by acting as heat sinks. These results indicate that no unanticipated pressure effects result from forced turbulence even at high concentrations of hydrogen.

The extended geometry tests were performed by attaching the pipe to the side of the sphere. The effects of varying ignitor location, fans, and unequal concentrations in each vessel were investigated. The hydrogen concentration varied between 6-25 volume percent. All of these tests were run without adding steam.

Results of varying the igniter locations between the end of the pipe and the conter of the sphere confirmed that lean mixtures propagate a fisme more readily in the upward than horizontal direction and in the presence of turbulence. Although the burst disc initially separating the mixtures in the pipe and sphere induced local turbulence which enhanced the rate and extent of reaction, no significant of fects of propagating fismes between unequal concentrations were observed. Even in a long, narrow pipe, at high concentrations of hydrogen with no steam present, no detonation occurred.

The Whitoshell tests investigated a number of parameters related to the potential hydrogen combustion phenomena inside the containment. Based on their results, we conclude that the GM and Tayco ignitors would reliably ignite lean mixtures of hydrogen in a postaccident environment. We also conclude that the observed i effects of steam, induced turbulence, connected geometries, and unequal concentrations on the nature of hydrogen combustion have confirmed our previous understanding. None of the resnits would prec1sde the application of distributed ignition for pcstaccident l

hydrogen control. In particular, the tests are important for what they did not show, the occurrence- of a detonation even in the presence of extremely severe conditions.

C. Water Micro-Foa Inertina - Factorv Mutual Research Cornoration The Factory Mutual project was the first of a two-part experimental program to investigate the pressure suppressant ,

offects of a water micro-fog. The purpose of the Factory Mutual proj ect was to experimentally identify in small scale a set of nominal micro-fos conditions for investigation in the Acurex

intermediate scale hydecasa combustics st:dios (Ssctica IV.D).

Since the interest was in the pressure suppressant effects of a water micro-fog, the Factosy Mutual proj ect was necessary in order to avoid inadvertently inerting the Acurer test vessel.

Therefore, the approach taken by Factory Mutual to achieve the project objective was to experimentally detemnine the water micro-fog requirements for inerting hydrogen-air mixtures and then simply recommend to Acurex a set of micro-fos conditions that did not meet those requirements. Emphasis was placed on visually dense fogs with number mean droplet sizes between 1-100 microns.

Tests were conducted in a plexiglas tube approximately 3.5 feet long with a 6 inch inner diameter. A 2.8 Joule spark served as the ignition source. Several tests were also conducted with a GM glow plus as the ignition source to verify the applicability of these tests to installed distributed ignition systems. The rm o-couples were used to determine the presence of combustion. Five different spray nozzles were used in order to obtain different fog conditions, i.e., a characteristic droplet size and density. Varying the pressure drop across each spray nozzle also allowed diff erent fos conditions to be obtained. Additionally, the micro-fog temperature and hydrogen concentration were varied.

Test results showed that at sabient conditions, visually dense ,

water micro-foss only marginally increase the hydrogen lower flammability limit. Additionally, as the characteristic droplet size is increased, the fog density required to maintain the same level of inerting is significantly increased. It was also demonstrated that increasing the micro-fog tanperature increases the effect on the hydrogen lower flammability limit. Finally, the Factory Mutual tests showed that a glor plus and a strong spark 4 source perfonned with no noticeable diff erence in combustion r e sul t s.

D. Hydromen Combustion Control Studies - Acurex Cornoration i

The Acurez project consisted of two phases. Phase 1 investigated l the effect of igniter location within an enclosed compartment, while Phase 2 was the second of the two-part water micro-fog program (see Section IV.C). Quiescent tests have been conducted by other organizations where the ignition source location was varied. How ev er, conditions inside the contaianent during a degraded core accident cannot be considered quiescent. Thus, the purpose of the Phase 1 test program was to qualitatively address l

the bsportance of igniter location during transient conditions.

l The purpose of the Phase 2 test program was to experimentally l

investigate the pressure suppressant effects of the two water micro-fog conditions recommended by Factory Mutual in both transient and quiescent tests.

Tests were conducted in a 17-foot high vessel with a 7-fogt inner dinneter. The total free volumo was approximately 630 f t Thermocouples were used to detect flame f ront location and vessel atmosphere temperature. Strain gauge and piezoelectric pressure transducers were used to measure the vessel abnospheric pressure.

Transient tests were conducted in Phases 1 and 2 with a continuous

irjostics of either hydregon er a hydregon-steam cinturo. Tho

. hydrogen and hydrogea-steam flow rates used in the tests were falculated by applying the volume ratio of the test vessel and the combined lower and ' dead-ended' plant compartments to the average release retos calculated with the MARCH Code for an 8 0 ***IO*"" 2 l sequence. An igniter assembly supplied by Duke Power was preenergized for all transient tests. In the Phase 1 tests, the l igniter was located either near the top, at the center, or near l the bottom of the test vessel. Some Phase 1 tests were conducted with water sprays present. Phase 2 tests were conducted both with and without two separate miere-fog conditions and with various hydrogen concentrations. The Phase 2 transient tests were ,

conducted with the bottom igniter location.

Results of the Phase 1 tests ladicated that ignitor location has same eff ect on combustion characteristics. This effect was shown to depend on: (1) whether the test was quitscent or transient, (2) the location of the igniter relative to the hydrogen source, and (3) the amount of turbulence present. The te sts showed that, during transiest inj ection periods, the pressure rise was less when the igniter was located near the region where the entering hydrogen mixed and first became flammable. The location of this region within oostainment would be determined by the geometry of each plant compartment, the hydrogen entry location and velocity, and the presemos of turbulence within the compartment. Since these tests have demonstrated the desirability of near-limit combustion, we conclude that ignitors should be located in the ice condenser upper plenum to allow near-limit combustion'to occur as the hydrogen exits from the ice condenser. The Phase 1 tests also indicated that the potential for a larger pressure rise existed when the hydrogen source j et continued to bypass the igniter until the bulk of the vessel had reaahed a flammable concentration.

This would tend to support locatlag ignitors in the upper portion of the lower compartment to preolude the source j et f rom i potentially bypassing nearby ignitors. It is important to note j that multiple ignitors were loosted throughout the containment i

regions at various elevations to ensure near-limit combustion (see Section II). In addition, it is noteworthy that the Banford tests (described in Section IV.B) demonstrated that the lower compartment region would be well-mised, which, according to the Acurez tests, tends to reduce the signifloanoe of igniter location relative to the inlet mixing region. The Phase 1 tests also confirmed previous findings on the pressure sitigative effects of steam and water sprays due to turbulence-induced mixing.

Results of the Phase 2 tests showed that a water micro-fog had no pressure sitigative ef fect during hydrogen combustion in quiescent mixtures. This indicated that the dominant effect of the fog droplets was not as a heat sink. The pressure sitigative effect of micro-foss in the transient tests seemed to be due to induced turbulence similar to the ef fect of sprays in some of the Phase 1 tests. This induced turbulence promoted mixing which enhanced the potential for near-limit combustion of the entering hydrogen.

I

Since an ice condenser containment would be suf ficiently ttrbulcat to ensure good mixing during a degraded core accident (see Section IV.E for a discussion of the Banford tests), we conclude that inducing additional turbulence with micro-fogging would be unne ce s sa ry.

In addition to the above conclusions based on the test obj ectives, an evaluation of the tests revealed additional information f rom which conclusions were drawn. The Gk igniter assemblies, identical to those in Duke Power's McGuire Nuclear Station and very similar to those used in the TVA IDIS, survived over five caulative hours of exposure to combustion test environments. The assembly and power cable continued to operate without f ailure, ne second additional conclusion dealt with estimated flame speeds. Although the test was not specifically instrumented to obtain flame speeds, it was possibib to calculate ' average' flame speeds f rom the pressure rise data of several transient and l quiescent toats. He calculate,d flame erseds in the transient tests varied f rom 1-2 f t/sec with steam present and either top or bottom ignition to 4 f t/sec with no steam present and bottom ignition. Fisme speeds f rom the quiescent tests varied f rom 3-8 f t/sec as the hydrogen concentration was increased from 5 to 11 volume percent. Thus, we conclude that these data support the flame speed ranges used in the 0,ASIX analyses (see Section III.A). Another important result of the transient test series was that the nature of combustion was always deflagrative instead of detonative even when a hydrogen-rich mixture was entering the vesse). Perhaps the most significant observation was the extreme contrast in pressure rise between quiescent and transient i combustion tests. The pressure rises during all of the transient tests in both Phase 1 and 2 was dramatically less than during the quiescent tests (with the exception of one very loan mixture

quiescent test). From this contrast, we conclude that caution must be used in the direct application of data from quiescent tests to the investigation of transient conditions. A final conclusion is that since the expected contalment postaccident environment would more closely resemble the transient test conditions, it follows that the pressure rises from sequential combustion should be relatively benign.

E. Hydromen Distribution - Hanford Enaineerina Develonnent Laboratory Tests were conducted at Hanford to investigate the potential for nonuniformities or gradients in the distribution of hydrogen during a degraded core accident in an ice condenser containment.

The purpose was twofold: (1) to investigate whether the potential existed f or pocketing of rich mixtures that could lead to a local detonation and (2) to determine whether the well-mixed nodalization assumptions in the contatsment analysis were valid.

The eff ects of temperature, forced circulation, and j ets were studied. The emphasis was placed on representing a small break LOCA in the ice condenser containment since that was tha base case used for design and analysis of the ignition system.

.u. -

The B::ferd Centsiancat Systems Test Fgoility can s31osted b20saso

. fts relatively large volume (30,000 f t ) reduced scaling effects '

and because its interior conid be customized to represent the structures of an ice condenser containment. Helium was used as a '

simulant for hydrogen in most of the tests due to site safety re gula tions.

Since the upper compartment of the ice condenser containment is well mixed by the sprays, the lower campartment region was chosen i for modeling emphasis in the f acility. A divider deck, reactor cavity, refueling canal, the air return f ans and ice condenser t

lower inlet doors were all represented. The hydrogen (helium)/steen release was scaled from small break LOCA calenistions using the MARCH computer code. Two release scenarios were modeled: (1) a 2' pipe break with a horizontal orientation and (2) a 10' pressuriser relief tank rapture disc opening with a vertically upward orientation. Atmospheric temperatures, velocities, and gas concentrations were acesared at several distributed sample points during the tests.

The test resnits showed that mixing was very good, even without

, forced circulation by the air return fans. The maximum hydrogen concentration difference at any time during the release between ,

any two sample points in the lower compartment was 2-3 volume percent. In addition, these concentration differences had stopped increasing even before the release period was over. We conclude that there is no potential for pocketing of rich mixtures and that the well-mixed assumptions in the containment analysis were j justified, i

F. Cable Survivability and Igniter Durability - TVA Singleton Materials E==ineerina Laboratorw Tests were condssted at Singleton to demonstrate the survivability of electrical cable and the durability of both GM and Toyoo ignitors. Samples of the exposed incore thermocouple and hot and cold leg RTD cables and the igniter assembly power cable in condait were subjected to temperatures conservatively higher than

, calculated containment atmospheric temperature profiles during hydrogen barns. In a separate test series, the GM and Tayco ignitors were subj ected to durability testing consisting of thermal cycling, endurance, and combustion.

Since surf ace temperature effects could be important to the  ;

survivability of exposed thermocouple and RTD cable in the

-contatmaont, tests were conducted at Singleton in lieu of analysis. A transient temperature profile that conservatively bounded the calculated transient atmospheric profile of the lower compartment (where the thermocouple and RTD cables are located) was imposed on the exposed cables in an oven. An indication of the conservatism of the test was the f act that the measurement thermocouple placed inside an outer cable j acket showed temperatures during the test even higher than the peak caloniated atmospheric temperature in containment. In another test, a constant temperature profile that conservatively bounded the integrated heat finx f rom the cales1sted transient atmospheric

. _ - _ . . _ . ~ . _ _ _ _ _ - - . . _ _ . _ _ _ _ , . . _ _ . . . _ , . _ _ _ . _ . . _ _ _ _ _ _ . . .

profile of the upper plenan (where the igniter powcr cablo oculd be exposed to the most burns) was imposed on the cable in conduit in an oven. The f act that the cable reached and maintained internal temperatures during the test well above the calculated cable temperature is evidence of the conservatism of this test.

Following earth of the tests, all the cable inssistion successfully passed visual inspection and a resistance cheek for breakdown under high voltage. We conclude that both the exposed cable and cable in ocaduit would survive a degraded core accident that included Isydrogen combustion.

, Durability tests were performed at Singleton on both the GM and Tayco ignitors. The thermal cycling tests consisted of repeated

activations in air at several constant voltages. The endurance

' tests consisted of activation at several constant voltages for extended periods of up to one week. The combustion tests consisted of activations in both a premixed closed vessel and in a flowing mixture in an open combustion tube. Each of the igniter types continued to operate satisf actorily during all of these tests and successfully passed posttest visual inspections. We conclude that either the GM or Tayco igniter is sufficiently durable to provide controlled ignition in a degraded core accident.

J f

4

{

l l

V. Conclusions TVA has designed a Permanent Hydrogen Mitigation System employing controlled ignition to mitigate the effects of hydrogen during potential degraded core accidents at the Sequoyah Nuclear Plant.

The system is redundant, capable of functioning in a postaccident environment, seismically supported, capable of actuation f rom the main control room, and has an ample number of ignitors distributed throughout the containment. The containment structures and key equipent have been shown by analysis or testing to survive the pressure and temperature loads from selected degraded core accidents and to continue to function. An extensive research program has confirmed our analytical assumptions, demonstrated equipent survivability and shown that controlled ignition can indeed mitigate the effects of hydrogen releases in closed vessels. We conclude that the PBMS is an adequate hydrogen control system that would perform its intended function in a manner that provides adequate rafety margins.

8 4

4

VJ.u ReferencesSection II

- Sequoyah Nuclear Plant Hydrogen Study, Volume I (letter f rom L. M. Mills to A. Schwencer dated September 2,1980)

- Second Quarterly Research Report ' letter from L. M. Mills to A. Schwencer dated March 16, 1981)

- 1hird Quarterly Research Report (letter from L. M. Mills to E. Adensen dated June 16, 1981)

- Selection of the Permanent Hydrogen Mitigation System for the Sequoyah Nuclear Plant (letter from M. R. Wisenburg to E. Adensam dated July 1,1981)

- Fourth Quarterly Research Report (letter from L. M. Mills to E. Adensam dated September 22, 1981)

- Response to Additional NRC Questions on Hydrogen Control System (letter f rom L. M. Mills to B. Adensam dated December 1,1981)

Section III.A

- Sequoyah Nuclear Plant Hydrogen Study, Volume II, Revision in Response to NRC Questions (letter from J. L. Cross to R. L. Tedesco dated December 11, 1980)

- Additional Inf ormation Requested by NRC (1stter from J. L. Cross to R. L. Tedesco dated December 17, 1980)

- Resolution of Equipment Survivability Issues for the Sequoyah Nuclear Plant (letters f rom L. M. Mills to B. Adensam dated June 2,1981, and June 3,1981)

- CLASI1 Topical Report (letter f rom L. M. Mills to E. Adensam dated December 1,1981)

- Response to Additional NRC Questions on Hydrogen Control System (letters from L. M. Mills to E. Adensam dated December 1,1981, and J annary 5,1982)

Section III.B

- Sequoyah Nuclear Plant Hydrogen Study, Volume II, Revision in Response to NRC Questions (letter from J. L. Cross to '

R. L. Tedesco dated December 11. 1980)

- Additional Inf ormation Requested by NRC (letter from J. L. Cross to R. L. Tedesco dated December 17, 1980)

- Resolution of Equipment Survivability Issues for the Sequoyah Nuclear Plant (letters f rom L. M. Mills to E. Adensas dated June 2,1981, and June 3,1981) t cy

- Responso to NRC Rcqcost fer Infcrmatice on Equipment

" ' Survivability f or Segnoyah (letter L. N. Mills to E. Adensam dated December 1,1981)

- Response to Additional NRC Questions on Hydrogen Control System (letter from L. N. Mills to B. Adensam dated December 1,1981)

Section IV. A

- Sognoyah Nuclear Plant Hydrogen Study, Volume II, Revision in Response to NRC Questions (letter from J. L. Cross to R. L. Tedeseo dated December 11, 1980)

- - Second Quarterly Research Report (letter from L. N. Mills to A. Schwencer dated March 16, 1981)

Section IV.B

- Fif th Quarterly Research Report (letter from L. N. Mills to E. Adensam dated January 22, 1982)

- Sixth Quarterly Research Report (letter f rom L. N. Mills to E. Adensam dated April 23, 1982)

- Summary of Testing to Determine Snitability of Tayco Igniter for Use in the Permanent Bydrogen Mitigation System at Segnoyah and Watts Bar Nuclear Plants (letter f rom L. M. Mills to E. Adensas dated June 14, 1982)

- Seventh Quarterly Research Report (letter from D. S. Kammer to E. Adensas dated July 28, 1982)

Section IV.C I

- Fif th Quarterly Research Report (letter from L. N. Mills to E. Adensas dated January 22, 1982)

Section IV.D

- Fif th Quarterly Research Report (letter f rom L. N. Mills to E. Adensam dated January 22, 1982)

Section IV.E

- Fif th Quarterly Research Report (letter from L. N. Mills to E. Adensam dated January 22, 1982)

Section IV.F

- Sognoyah Nuclear Plant Hydrogen Study, Volume II (letter f rom L. M. Mills to A. Schwencer dated September 2,1980)

- Segnoyah Nuclear Plant Rydrogen Study, Volume II, Revision in Response to NRC Questions (letter from J. L. Cross to R. L. Tedesco dated December 11, 1980) w- w w , w yes v w e- - - .------ --~.- - - - - ----,w r-, - -w-- - e---- -m- - - - - - - - - - - - - - ~ - - - - ,- - - - - - - - - < - - , - - - -- , ,

I

, ' 'First Quarterly Research Report (letter from J. L. Cross to l R. L. Tedesco dated December 11, 1980) l

- Second Quarterly Research Report (letter from L. M. Mills to A. Schwencer dated March 16, 1981)

- Resolution of Equipment Survivability Issues f or the Segnoyah Nuclear Plant (letters from L. M. Mills to E. Adenssa dated June 2,1981, and June 3,1981)

- Fourth Quarterly Research Report (letter from L. M. Mills to E. Adensas dated September 22, 1981)

- Response to Additional NRC Questions on Hydrogen Control System (letter f rom L. M. Mills to E. Adensam dated December 1,1981)

- Response to NRC Request f or Information on Equipment Survivability for Sequoyah (letter f rom L. M. Mills to E.

Adensam ' dated December 1,1981) l

- , - -